Analysis of Crash Risk Perception Factors in Urban Road Network Using GPS Data and Sensitivity Analysis for Stop Speed Threshold

Abstract

Two-fluid Model, which describes traffic flow in urban road network macroscopically, was mainly used for operational analysis of network. Recently, the studies on driver’s behavior and safety using Two-fluid Model were conducted. Especially, the estimation of the crash risk perception factors was analytically suggested based on the model. Accordingly, this study aimed to validate from the real data that the safety of urban road network could be macroscopically analyzed with the crash risk perception factors. The factors were estimated from digital tachograph data of taxis in Jongno, Seoul and Gangnam, Seoul, and the estimates showed that Gangnam with more average lanes had higher crash probability and Gangnam with less intersections had lower crash severity than Jongno. The results were consistent with the results from the previous studies with other methodologies. Meanwhile, stop speed threshold, which is introduced to accommodate the error in GPS data like digital tachograph, affects the estimation of Two-fluid Model parameters. The crash risk perception factors were also expected to vary with the stop speed threshold, and sensitivity analysis was therefore conducted. As a result, the factors increased with stop speed threshold increasing. In addition, the two factors fluctuated similarly, which implies the existence of the consistent relationship between them. This study estimated the crash risk perception factors of individual regions from taxi digital tachograph data, and verified the possibility of the macroscopic safety analysis in urban road network. The macroscopic studies on the crash risk in urban road network are anticipated in the future based on this study.